A groundbreaking online reference library improves accuracy in identifying drug exposure
Medication histories often rely on patient recall or medical records, but both can miss important details. People may forget what they took, use over‑the‑counter drugs, keep leftover prescriptions, buy medicines online, or encounter drugs indirectly through food and the environment. These gaps mean that significant drug exposures can go undetected, yet knowing which drugs are present can have major implications for health and biology.
Researchers from the University of California San Diego and colleagues have created a publicly accessible online reference library of chemical fingerprints for thousands of drugs, their breakdown products, and related compounds. Published in Nature Communications on December 9, 2025, this GNPS Drug Library enables more precise detection of drug exposure by comparing unknown compounds in a patient’s blood, urine, or other biological samples to the library’s entries.
The library was built using mass spectrometry, which charges molecules to determine their weight and then fragments them to generate distinctive chemical fingerprints. Each drug entry includes information about its source (prescription, over‑the‑counter, etc.), its drug class, its typical uses, and how it acts in the body.
To assess the library’s effectiveness, the team employed untargeted metabolomics, a mass spectrometry approach that analyzes thousands of molecules at once to identify drug metabolites within a sample. In this framework, any sample—whether urine, breast milk, or even environmental water—can yield a comprehensive snapshot of its chemical contents.
Nina Zhao, Ph.D., co‑first author, notes that the approach confirmed many expectations: commonly prescribed drugs emerged in patterns consistent with diagnoses such as inflammatory bowel disease, Kawasaki disease, and dental caries, while antifungal agents appeared in skin‑related samples from people with psoriasis.
The library’s utility was further tested with nearly 2,000 participants from the American Gut Project, which studies gut microbial diversity across the United States, Europe, and Australia. This analysis identified 75 distinct drugs, reflecting the most commonly prescribed drug classes and medications in those regions.
Key findings include:
- U.S. participants tended to show more detectable drugs per person than European or Australian participants.
- Pain relievers appeared more often in female samples, while erectile‑dysfunction medications were more commonly detected in male samples.
- The library can reveal clinically relevant co‑existing medications in conditions such as Alzheimer’s disease, where cardiovascular and psychiatric drugs were detected, aligning with common comorbidities.
- In HIV patients, antiviral medications were found alongside cardiovascular and psychiatric drugs, consistent with higher rates of heart disease and depression in this group. The data also linked certain HIV meds to specific changes in gut‑derived molecules, illustrating how drug exposure can influence the microbiome.
A central aim is to support precision medicine by explaining why patients respond differently to treatments based on how their bodies metabolize drugs. Zhao emphasizes that understanding gut microbiome interactions with various medications could help tailor therapies more effectively.
Beyond human samples, the library revealed that antibiotics appear in meat products and pesticides used in vegetables—indicating environmental exposures that can impact human health. The researchers even found antibiotics in food and environmental samples like reclaimed water and snow, suggesting broader sources of drug exposure that the library can help uncover.
The GNPS Drug Library sets a new standard for linking drug exposure, microbial metabolites, and patient outcomes. It will continue to grow, with ongoing exploration of large language models and generative AI to curate new data. A user‑friendly online data analysis app allows researchers without pharmacy backgrounds to quickly analyze their datasets, producing drug identifications, metrics, and visualizations with a single click.
The broader vision is to enable clinicians and researchers to understand why certain patients do not respond to treatments as expected, by revealing the precise medication exposures driving biology and health outcomes. Ultimately, this resource could help optimize drug therapy and advance personalized medicine.
Source:
Zhao, H. N., et al. (2025). A resource to empirically establish drug exposure records directly from untargeted metabolomics data. Nature Communications. doi: 10.1038/s41467-025-65993-5.
